Connected objects and big data have both become essential in order to engage with and remain in contact with consumers, and to boost sales. But in these troubled times, how can they protect us? An explanation from Jean-Philippe Cunniet.
New technologies must play a role to protect the population at three key moments:
BEFORE: it is possible to improve predictions of crimes and attacks before they take place.
DURING: improve the reaction to attacks, at the time and in the place they occur.
AFTER: find the culprits and speed up investigations and arrests.
Algorithms based on “Big Data + connected objects” will be more effective, but they require connected sensors to be installed in all our towns and cities.
Technologies can prevent crimes BEFORE they take place.
In the film Minority Report (2002), the crimes were predicted and prevented before they even took place, thanks to technology. In the United States, Palantir and Forensic Logic, two companies from California, predict terrorist crimes on the basis of the data available on all the terrorist acts in the past. Similarly, Future Crimes Institute brings together technical experts from the world of security in order to help police forces to understand the risks and imagine the solutions.
The Kaggle big data platform recently enabled some of the best computer scientists in the world to analyse a set of security data from the city of San Francisco in order to develop this type of predictive algorithm. Crimes that are not premeditated, such as crimes of passion, pass through the mesh of the algorithms, but according to the former police officer and Interpol consultant, Goodman, “for several years, the analysis of big data has enabled us to predict numerous premeditated crimes, including car theft, street crime, other thefts and homicides”.
When “a number comes up” in the TV series “Person of Interest” (or from a computer algorithm), the police must decide how to use this information. Should they ask a magistrate to reinforce the surveillance of a person who is “only” designated by an algorithm? In the absence of any proof, the police must find the just means of using these statistics, without infringing the rights of innocent citizens.
The data does not always show the whole situation, and the algorithms take decisions according to their configuration at a given time… Street lighting is one of the best preventive measures. The better the lighting, the safer the street. So should drones be used to light up sensitive zones at certain times by analysing big data on crime risks? The data may include the genetic code of suspects, information on facial recognition or suspect withdrawals from cash dispensers equipped with cameras, which are often used by potential terrorists to collect cash from other countries.
On the one hand, the Bitcoin is an effective means of remaining anonymous in monetary transactions, just like “Tor” can be used to buy weapons anonymously. On the other hand, the French terrorist Coulibaly took out a loan with Cofidis to fund his attack.
Technologies can prevent crimes WHILE they are taking place.
The streets of major cities all over the world will probably be equipped with even more CCTV cameras (including thermal and night vision cameras), because they are so useful in police investigations. These devices will be accepted (or even demanded) by the population, depending on the degree of insecurity. The cameras supplied by the American company FLIR are currently amongst the most efficient. But these technologies must be used to protect public places without infringing individual freedoms.
Security gates can only be installed at the entrance to public places if the public is informed (according to the rules of the French agency for Information Technology and Individual Freedom). And the public must be able to pass through the gate “without noticing it and without being slowed down”, so that they do not feel that the world is not as safe as it used to be. The controls must be “invisible and fluid”.
The long-established players on the security stage, such as Morpho Detection have a role to play. Why are there still no revolving gates equipped with metal or explosives detectors? This type of detection solution “looks natural” and would make the population feel less anxious. CCTV images can be analysed by algorithms, by employees or by the population. So why not anonymously broadcast the images from CCTV cameras on the internet, with regular switches between cameras, so that everyone can collaboratively monitor their own town, country or planet, according to the principle that states “If you control my streets, then I control yours”.
This principle was introduced for the scanned images of baggage in airports, but it came up against opposition from the general public. Luggage checks have not yet benefited from the emergence of the collaborative economy, to the detriment of security. Are the French prepared to accept the principle of “crowd watching” in CCTV images? There is no point in complaining about “Big Brother” , because citizens will regain control of the images of the life in their streets, which are already published by definition.
Sound intensity detectors can automatically recognise particular noises, such as gunfire, and then automatically send an alert to the forces of law and order. This means that an individual citizen, who may be faced with a weapon of war, does not have to raise the alert.
Detection of explosives
Can sniffer dogs be replaced by connected objects, an invisible and non-intrusive solution that is also used to measure air quality in public places?
Everyone now has a smartphone, some of which are capable of recording fingerprints. But smartphones can be hacked and falsified, so they are not the ideal solution.
Certain police operations could be performed by drones, fitted out with particular equipment for the mission:
– cameras to take pictures of assailants for high-speed facial recognition,
– onboard weapons to stop criminals, such as tear gas launchers that immobilise assailants,
– dazzle drones.
Ambulance drones Already tested in the Netherlands in order to provide first aid and save time in the rescue of as many victims as possible, with the help of personnel that is already on the spot. A French company also proposed ambulance motorbikes that can weave through the traffic and deliver first aid, but this solution was never adopted, much to the detriment of the victims.
Far from the coercive electronic bracelet, geolocalisation technologies have grown strongly with the advent of the smartphone. Should every S-listed person be permanently geolocalised? This question raises the issues of individual freedom and rights that demands a strict legal framework. In this way, it will be possible to detect any suspicious and unusual movements, when certain individuals meet up.
The geolocalisation of suspects’ smartphones can already be permanently monitored in order to detect any unusual movements. What is an unusual movement? What are unusual meetings between people who are tracked by their smartphones? These traces are left by all the connected objects used by people who are under surveillance.
Geolocalisation of vehicles
Should every car be geolocalised in order to find, or even automatically stop, any stolen vehicle? Automatic number plate analysis video technology can also be used to check the movements of suspect vehicles (like the system used in London for the payment of tolls charged to enter the city).
The sharing of information between police forces is also essential. This is only possible in Europe if a common European governance policy and a European interior ministry is introduced. I hope that this will happen quickly in the years to come. How can we follow suspects from one city to another, from one region to another and from country to another, if we do not share data? Terrorism is global. Some of the terrorists who committed the attacks in Paris on 13 November were in Belgium on 12 November.
The Criminality Department of the ’University of Pennsylvania in the United States has developed its own algorithm to predict the next victims of homicides on the basis of a broad variety of data. Social networks also provide a lot of data to be analysed. Algorithms can now detect behaviours, key words and suspect exchanges. On 12 November, the account _jacky_boy_ posted on Instagram (an amateur joke or just a lie?) “ By the time you see this photo on BFM, it will be too late. ”
Technologies to manage the situation AFTER the crime
The arrival of the emergency services
In some countries, the fire brigades and the police can remotely control all the traffic lights on their route, so that they automatically turn green when they arrive. In a few years, the advent of automatic vehicles could make it much easier for the emergency services to arrive on the scene, by sending a message to all the cars in the vicinity to change their route.
Crimes are often committed by known offenders
Should algorithms be used to help magistrates to assess the risk of releasing, or not releasing, a prisoner, before the end of their sentence? In the United States, this is already possible in more than 80% of cases, by systematically analysing twenty-four items in the prisoner’s file, and the age at which the crimes were committed. This method has cut repeated offences by 15%.
Tracking after release from prison
Should everyone who is released from prison be condemned to registering all their movements, simply so that that they can be used in another investigation? I can already hear the voices of the champions of individual freedoms, and I agree with them. But this is technically possible. The boundary between individual freedom and security is very thin, and technology can be used to shift the cursor to one side or the other. I hope that the decision on the position of the cursor will come from the population itself.